Locally optimum and suboptimum detector performance in non-Gaussian noise
Abstract
The performance of adaptive locally optimum Bayes detectors (LOBD), which approach optimality under a vanishingly small signal situation, is compared with that of the commonly used, much simpler ad-hoc nonlinear detection schemes. The derivation of the LOBD for the simplest case of binary coherent phase shift keying is reviewed, and its performance is evaluated together with that of the hard-limiter by assuming that the desired signal becomes vanishingly small and that the time-bandwidth product is large. The results appear to indicate that the hard-limiter is always very close in performance to the LOBD for Class B impulsive noise. A situation where the hard-limiter outperforms the LOBD is demonstrated for one typical noise case. Thus, for Class B noise, it is probably not worthwhile to implement anything more complicated than a hard-limiter.
- Publication:
-
ICC 1982 - The Digital Revolution, Volume 1
- Pub Date:
- 1982
- Bibcode:
- 1982icc.....1R...2S
- Keywords:
-
- Bayes Theorem;
- Optimization;
- Performance Prediction;
- Random Noise;
- Signal Detectors;
- Computerized Simulation;
- Noise Reduction;
- Noise Spectra;
- Signal To Noise Ratios;
- Electronics and Electrical Engineering